A classification of alcoholics has been developed in a VA funded pilot study that is intended to meet four criteria for a good taxonomy. It is hoped that this classification, bacause it is based upon prognostic indicators, may be useful as a guide for type-specific treatment. In the present NIAAA study we are evaluating the classifications for interrater reliability and are comparing patterns in Duluth with those in Southern Wisconsin to see if classifications differ from region to region. The classification has not yet been evaluated for its effectiveness as a guide to type-specific treatment; so no well substantiated claim can be made about its utility at this time.